Evaluating the Optimality of Dynamic Coupling Strategies in Interdependent Network Systems

Cascading failures are a common phenomenon in complex networked systems, where failures at only a few nodes may trigger a process of sequential failures. We investigate the robustness against cascading failures in systems carrying flows or loads that contain multiple interdependent networks, e.g., power grid, transportation system, etc. In these systems, the coupling coefficients between the networks, which determine how the flow from failed components gets redistributed across the networks, is a key factor affecting the robustness against cascading failures. Prior work has introduced the step-wise optimization (SWO) strategy that dynamically adjusts the coupling coefficients during the course of the cascading failures in an effort to preserve the network size. SWO has been shown to have good performance against cascading failures on synthetic data. In this paper, we show the optimality of the SWO strategy under certain conditions on the flow and capacity distributions of the nodes. We also show, via simulations, that the SWO strategy performs well under various real-world network topologies as well.